“Make this sound better” is probably one of the most common prompts used in Claude, ChatGPT, and Gemini, but I’m worried it’s turning us all into carbon copies of each other, especially at work.
To test my theory, I used Gemini on my work emails for a week to see how it performed. It was certainly efficient, but after a week, I felt like I'd completely lost all my personality.
With over 75% of professionals now using AI tools to write or refine workplace communication, the shift is no longer subtle.
Dan Bruce, founder of PressReacher, thinks this growing “AI Personality Shift” could be changing workplace culture. “Over time, that starts to blur the line between their natural, normal voice to an AI-enhanced one”, he says.
AI writing tends to follow the same pattern: neutral, polished, easy to read, and noticeably less emotional, and it is becoming part of our work culture.
AI delivers a level of polish and politeness to your emails that makes them feel a bit unnatural and generic. I signed off everything for a week with “Kind regards”, and after a few days, I felt a part of my soul dying.
5 tips for using AI and keeping your personality
If you want to use AI for work email but don't want to lose your voice entirely, Bruce offers some tips for how to keep using AI while avoiding ‘AI Personality Shift’.
1. Draft first, then refine: “Write your message before turning to AI, not the other way around”, recommends Bruce. This is probably the key tip if you want to retain your own voice, but still use AI. If you ask AI to write something for you, then you’ve already lost your voice.
2. Keep your tone and quirks: “Don’t remove personality completely. Natural phrasing builds better connections”, he advises. Again, this is another good tip. If AI tries to smooth out the parts of your email that make you, er, you, then shut it down.
3. Use AI as a tool, not a voice: “Think of it as an editor or a proofreader, not a replacement”, says Bruce. I find a good tip here is to ignore a lot of its proofreading advice - correcting grammatical errors is fine, but most of the time, AI wants to make “smoothing” changes that completely rip your personality out of any text.
4. Avoid overdoing everyday messages: “Not every Slack or Teams message needs to sound perfect”, reminds Bruce. There’s no need to use AI for every little thing.
5. Sense-check, don’t default: If you’re using AI, it becomes very tempting to turn your brain off and default to it as an authority, when in fact it’s a better servant than a master. “Ask yourself, ‘Does this sound like me?’”, suggests Bruce. If it doesn’t, you know what to do.
Subtly flattened
Business communication requires a certain degree of formality, I get that, but I worry we’re having our individual personalities subtly flattened by AI when we use it in the workplace.
I think it’s the diversity of voice and opinion that makes a team perform better when they work together. If we all end up as homogenised versions of each other, we risk losing one of our chief assets: our individuality.
After a week of AI emails, I’m done with it. If everything sounds perfect, but not like me, then what exactly am I improving?
